Chee Shiu Chung = 阻尼振子的線空間結構與哈密頓理論 / 朱兆中. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2003. / Includes bibliographical references (leaves 88). / Text in English; abstracts in English and Chinese. / Chee Shiu Chung = Zu ni zhen zi de xian kong jian jie gou yu ha mi dun li lun / Zhu Zhaozhong. / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Conservative Systems --- p.4 / Chapter 2.1 --- General Formalism --- p.4 / Chapter 2.2 --- One Simple Harmonic Oscillator --- p.7 / Chapter 2.3 --- Two Coupled Harmonic Oscillators --- p.9 / Chapter 3 --- Dissipative Systems --- p.12 / Chapter 3.1 --- Elimination of Bath --- p.12 / Chapter 3.2 --- One Oscillator with Dissipation --- p.16 / Chapter 3.3 --- Two Oscillators with Dissipation --- p.19 / Chapter 4 --- Eigenvector Expansion and Bilinear Map --- p.21 / Chapter 4.1 --- Formalism --- p.21 / Chapter 4.2 --- Inner Product and Bilinear Map --- p.23 / Chapter 4.3 --- Normalization and Phase --- p.25 / Chapter 4.4 --- Matrix Representation --- p.25 / Chapter 4.5 --- Duality --- p.28 / Chapter 5 --- Applications and Examples of Eigenvector Expansion --- p.31 / Chapter 5.1 --- Single Oscillator --- p.31 / Chapter 5.2 --- Two Oscillators --- p.32 / Chapter 5.3 --- Uneven Damping --- p.33 / Chapter 6 --- Time Evolution --- p.36 / Chapter 6.1 --- Initial-Value Problem --- p.36 / Chapter 6.1.1 --- Green's Function --- p.37 / Chapter 6.2 --- Sum Rules --- p.39 / Chapter 7 --- Time-Independent Perturbation Theory --- p.41 / Chapter 7.1 --- Non-degenerate Perturbation --- p.41 / Chapter 7.2 --- Degenerate Perturbation Theory --- p.46 / Chapter 8 --- Jordan Block --- p.48 / Chapter 8.1 --- Jordan Normal Basis --- p.48 / Chapter 8.1.1 --- Construction of Basis Vectors --- p.48 / Chapter 8.1.2 --- Bilinear Map --- p.50 / Chapter 8.1.3 --- Example of Jordan Normal Basis --- p.55 / Chapter 8.2 --- Time Evolution --- p.56 / Chapter 8.2.1 --- Time Dependence of Basis Vectors --- p.56 / Chapter 8.2.2 --- Initial-Value Problem --- p.58 / Chapter 8.2.3 --- Green's Function --- p.59 / Chapter 8.2.4 --- Sum Rules --- p.60 / Chapter 8.3 --- Jordan Block Perturbation Theory --- p.61 / Chapter 8.3.1 --- Lowest Order Perturbation --- p.61 / Chapter 8.3.2 --- Higher-Order Perturbation --- p.65 / Chapter 8.3.3 --- Non-generic Perturbations --- p.66 / Chapter 8.4 --- Examples of High-Order Criticality --- p.66 / Chapter 8.4.1 --- Fourth-order JB --- p.67 / Chapter 8.4.2 --- Third-order JB --- p.74 / Chapter 8.4.3 --- Two Second-order JB --- p.79 / Chapter 9 --- Conclusion --- p.81 / Chapter A --- Appendix --- p.83 / Chapter A.l --- Fourier Transform and Contour Integration --- p.83 / Chapter B --- Degeneracy and Criticality --- p.86 / Bibliography --- p.88

Fan, Sin Tsun Edward. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2009. / Includes bibliographical references (leaves 85-87). / Abstract also in Chinese. / Chapter 1 --- Introduction --- p.5 / Chapter 1.1 --- Motivation --- p.5 / Chapter 1.2 --- Main results --- p.10 / Chapter 2 --- Preliminaries --- p.14 / Chapter 2.1 --- Z-gradations of Semisimple Lie Algebras --- p.14 / Chapter 2.2 --- Basic Facts about Algebraic Groups --- p.15 / Chapter 3 --- Weight Multiplicity Free Representations and Pre- homogeneous Vector Spaces --- p.18 / Chapter 3.1 --- Weight Multiplicity Free Representations --- p.18 / Chapter 3.2 --- Prehomogeneous Vector Spaces --- p.22 / Chapter 4 --- Augmentations of Dynkin Diagrams --- p.25 / Chapter 5 --- Orbit Finiteness and Prehomogeneity --- p.32 / Chapter 6 --- Termination of Z-grading --- p.36 / Chapter 7 --- Explicit Construction of Generic Elements in Simply- laced Cases --- p.42 / Chapter 8 --- The Ambient Lie Algebras of Parabolic PVS's --- p.47 / Chapter 9 --- PVS's of Twisted Affine Type --- p.52 / Chapter 10 --- "Orbit Structure of (GL2 x SL2m+1,C2 x A2C2m+1)" --- p.55 / Chapter 11 --- Nilvarieties and Generalisation of Open Orbits --- p.59 / Chapter 11.1 --- Nilvarieties and Visible Representations --- p.59 / Chapter 11.2 --- Augmeantations of Affine Dynkin Diagrams --- p.62 / Chapter 11.3 --- Classification of Irreducible Visible Representations --- p.67 / Chapter 12 --- Real Forms of PVS of Parabolic Type --- p.70 / Chapter 12.1 --- Representations of Real Reductive Lie Algebras and Satake Diagrams --- p.70 / Chapter 12.2 --- Real Forms of PVS of Parabolic Type --- p.77 / Chapter 13 --- Tables --- p.81 / Bibliography --- p.85

Student Number : 0318930A - PhD thesis - School of Animal, Plant and Environmental Studies - Faculty of Science / The widespread use of pyrethroid insecticides has led to the emergence of significant insecticide resistance in various parts of the world. An unprecedented increase in the number of annual malaria cases reported in Kwazulu Natal, South Africa in 1999 to 2000 was attributed to the re-emergence of pyrethroid-resistant Anopheles funestus Giles. Resistance was metabolic-based with increased monooxygenase (P450) metabolising the pyrethroid insecticides. This emphased the need to understand the molecular mechanisms conferring pyrethroid resistance in An. funestus. The present study aimed to firstly isolate P450 genes in An. funestus and secondly, to identify P450 gene over-expressed in a resistant (FUMOZ-R) strain compared to a susceptible (FANG) strain. A third aim was to construct an An. funestus cDNA library to lay the foundation for future studies on P450 monooxygenses. Degenerate primers based on conserved regions of three An. gambiae P450 families were used to amplify cDNAs from An. funestus. Eleven CYP4, four CYP6 and five CYP9 partial genes were isolated and sequenced. BLAST results revealed that An. funestus P450s have a high sequence similarity to An. gambiae with above 75% identity at the amino acid level. The exception was CYP9J14. The An. gambiae P450 with the closest similarity to CYP9J14 exhibited only 55% identity suggesting a recent duplication event in CYP9J14. Molecular phylogenetic analysis also supported this hypothesis. Intron positions were highly conserved between the two species. Expression studies using blot analysis implicated CYP6P9, an ortholog of CYP6P3 in An. gambiae, as the over-expressed P450. Dot blot analysis revealed a 500-fold expression higher in FUMOZ-R strain compared with FANG strain. Semiquantitative PCR revealed that CYP6P9 was developmentally regulated. Expression was not detected in eggs and was higher in larvae compared to pupae. Quantitative real time PCR showed that CYP6P9 expression was 4.5-fold higher in 3-day old FUMOZ-R males than females and 3.5-fold higher in the 14-day old males than 14- day old females. Statistically, this difference was not significant suggesting that CYP6P9 expression is not sex specific. The An. funestus cDNA library construction in λTriplEx2 vector was successful with a titre of 4.9 x108 pfu/ml and a transformation efficiency of 98%.

Dans l’analyse des séries temporelles multivariées, la plupart des modèles existants sont utilisés à des fins de prévision, c’est-à-dire pour estimer les valeurs futures du système étudié à partir d’un historique de valeurs observées dans le passé. Une autre tâche consiste à extraire des causalités entre les variables d’un système dynamique. C’est pour ce dernier problème à visée explicative que nous développons une série d’outils. À cette fin, nous définissons dans cette thèse une nouvelle famille de modèles autorégressifs vectoriels non paramétriques construits à partir de noyaux à valeur opérateur. En faisant l’hypothèse d’une structure sous-jacente creuse, la parcimonie du modèle est contrôlée en imposant dans la fonction de coût des contraintes de parcimonie aux paramètres du modèle (qui sont en l’occurrence des vecteurs qui pondèrent une combinaison linéaire de noyaux). Les noyaux étudiés possèdent parfois des hyperparamètres qui doivent être appris selon la nature du problème considéré. Lorsque des hypothèses de travail ou des connaissances expertes permettent de fixer les paramètres du noyau, le problème d’apprentissage se réduit à la seule estimation des paramètres du modèle. Pour optimiser la fonction de coût correspondante, nous développons un algorithme proximal. A contrario, lorsqu’aucune hypothèse relative aux variables n’est disponible, les paramètres de certains noyaux ne peuvent être fixés a priori. Il est alors nécessaire d’apprendre conjointement les paramètres du modèle et ceux du noyau. Pour cela, nous faisons appel à un schéma d’optimisation alterné qui met en jeu des méthodes proximales. Nous proposons ensuite d’extraire un estimateur de la matrice d’adjacence encodant le réseau causal sous-jacent en calculant une statistique des matrices jacobiennes instantanées. Dans le cas de la grande dimension, c’est-à-dire un nombre insuffisant de données par rapport au nombre de variables, nous mettons en oeuvre une approche d’ensemble qui partage des caractéristiques du boosting et des forêts aléatoires. Afin de démontrer l’efficacité de nos modèles, nous les appliquons à deux jeux de données : des données simulées à partir de réseaux de régulation génique et des données réelles sur le climat. / In multivariate time series analysis, existing models are often used for forecasting, i.e. estimating future values of the observed system based on previously observed values. Another purpose is to find causal relationships among a set of state variables within a dynamical system. We focus on the latter and develop tools in order to address this problem. In this thesis, we define a new family of nonparametric vector autoregressive models based on operator-valued kernels. Assuming a sparse underlying structure, we control the model’s sparsity by defining a loss function that includes sparsity-inducing penalties on the model parameters (which are basis vectors within a linear combination of kernels). The selected kernels sometimes involve hyperparameters that may need to be learned depending on the nature of the problem. On the one hand, when expert knowledge or working assumptions allow presetting the parameters of the kernel, the learning problem boils down to estimating only the model parameters. To optimize the corresponding loss function, we develop a proximal algorithm. On the other hand, when no prior knowledge is available, some other kernels may exhibit unknown parameters. Consequently, this leads to the joint learning of the kernel parameters in addition to the model parameters. We thus resort to an alternate optimization scheme which involves proximal methods. Subsequently, we propose to build an estimate of the adjacency matrix coding for the underlying causal network by computing a function of the instantaneous Jacobian matrices. In a high-dimensional setting, i.e. insufficient amount of data compared to the number of variables, we design an ensemble methodology that shares features of boosting and random forests. In order to emphasize the performance of the developed models, we apply them on two tracks : simulated data from gene regulatory networks and real climate data.

Gene therapy is an attractive treatment for many genetic diseases because rather than treat the symptoms of the disease, it has the potential to correct the underlying defect. Cystic fibrosis and hemophilia A are two monogenic disorders that are particularly well-suited to treatment with gene therapy as a relatively small increase in the function is needed to see improvement. Gene therapy has provided some correction in both diseases using a variety of vector systems but sustained expression and long term correction have yet to be demonstrated in the clinic. It is unclear in which cell type(s) correction of the underlying defect in cystic fibrosis will be most effective. Studies indicate that the majority of CFTR expression is in the submucosal glands and ciliated epithelia – a terminally differentiated cell type (Engelhardt, J.F. et al, 2004, Journal of Clinical Investigation). Therapeutic gene transfer would thus be most effective if achieved in a progenitor cell type. Additionally, the native regulation of CFTR has not been definitively elucidated. To this end, one goal of our studies is to develop a lentiviral vector system with heterologous promoters of varying strengths and cell specificity to aid in our selection of optimal reagents for appropriate CFTR expression. We show that use of novel internal promoters from the human PLUNC and WDR65 genes direct persistent expression in the airway. Additionally, disruption of the nasal epithelium with the detergent polidocanol eliminated reporter expression in mouse airway. Two weeks post-treatment, expression returned indicating targeting of a progenitor cell population with our novel vectors. Integrating vector systems can treat chronic diseases such as cystic fibrosis because expression can persist long term from these vectors if cells with progenitor capacity are targeted (Sinn, P.L. et al, 2005, Journal of Virology). However, the potential for genotoxicity from vector-related dysregulation is a concern. Thus, a second aim of these studies was to develop a lentiviral vector that can target a specific locus in the genome. We developed a FIV vector in which the integrase was modified with a protein-binding domain that when co-delivered with a fusion consisting of the cognate protein and a DNA binding domain would tether the vector to the appropriate locus. Unfortunately, integrase modification rendered the vector catalytically inactive. Lastly, we hoped to develop a non-viral transposon vector system (piggyBac) for gene transfer applications to the liver for treatment of hemophilia A. The recent demonstration that piggyBac transposase is highly active in mammalian cells warrants further development of this vector as an alternative to other non-viral integrating vector systems currently under investigation. We showed persistent reporter and therapeutic transgene expression in the livers of mice treated with the piggyBac vector. Furthermore, we show for the first time in vivo persistence and increased expression from the recently developed hyperactive transposase. The development of integrating vectors targeted to specific tissues or genomic loci is important in for treatment of the monogenic diseases cystic fibrosis and hemophilia A.

Modulation of corneal transplant rejection using gene therapy shows promise in experimental models but the most appropriate vector for gene transfer is yet to be determined. The overarching aim of the thesis was to evaluate the potential of a lentiviral vector for use in human corneal transplantation. Specific aims were: (i) to assess the ability of an HIV-1-based lentiviral vector to mediate expression of the enhanced yellow fluorescent protein (eYFP), and a model secreted protein interleukin-10 (IL10), in ovine and human corneal endothelium; and (ii) to examine the influence of lentivirus-mediated IL10 expression on the survival of ovine corneal allografts. Four lentiviral vectors expressing eYFP under the control of different promoters, were tested: the simian virus type-40 (SV40) early promoter, the phosphoglycerate kinase (PGK) promoter, the elongation factor-1alpha (EF) promoter, and the cytomegalovirus (CMV) promoter. Two lentiviral vectors expressing IL10 were tested: one containing the SV40 promoter and another containing a steroid-inducible promoter (GRE5). Lentivirus-mediated expression in transduced ovine and human corneal endothelium was assessed by fluorescence microscopy, real-time quantitative RT-PCR and ELISA, following alterations of transduction period duration (2–24 hr) and vector dose, as well as in the presence or absence of polybrene or dexamethasone (GRE5 vector). It was also compared to expression mediated by adenoviral vectors. Orthotopic transplantation of ex vivo transduced donor corneas was performed in outbred sheep. Allografts were reviewed daily for vascularisation and signs of immunological rejection. Lentivirus-mediated eYFP expression was delayed in ovine corneal endothelium compared to human. However, in both species the final transduction rate was greater than 80% and expression was stable for at least 14 d in vitro. Lentivirus-mediated expression in ovine and human corneal endothelium was higher with the viral promoters in comparison to the mammalian promoters. A 24 h transduction of ovine corneal endothelium with the lentiviral vector encoding IL10 resulted in expression levels which were increasing after 15 d of organ culture but logarithmically lower than those achieved by adenovirus. Shortening the lentiviral transduction period to 2 h led to a reduction in expression, but the addition of polybrene (40 micrograms / ml) to the transduction mixture restored expression to levels comparable to those attained after a 24 h transduction period. Lentivirus-mediated IL10 expression was higher and more rapid in human corneal endothelium compared to ovine corneas. Dexamethasone-responsive transgene expression was observed in both ovine and human corneal endothelium using the lentiviral vector containing the GRE5 promoter. Lentivirus-mediated expression in ovine corneal endothelium was stable for 28 d in vivo. A modest prolongation of ovine corneal allograft survival (median of 7 d) was achieved by transduction of donor corneas for 2–3 h with the lentivirus expressing IL10. Attempts to increase the expression of IL10 by the addition of polybrene (40 micrograms / ml) to the transduction mixture, resulted in a toxic effect on corneal allografts which abrogated the beneficial effect of IL10. The lentiviral vector shows potential for the stable expression of therapeutic transgenes in human corneal transplantation. However, the mechanisms underlying the species-specific differences in HIV-1-mediated transgene expression will need to be elucidated and overcome if the ovine preclinical model is to provide justification for a clinical trial.

<p>Face detection is a preprocessing step for face recognition algorithms. It is the localization of face/faces in an image or image sequence. Once the face(s) are localized, other computer vision algorithms such as face recognition, image compression, camera auto focusing etc are </p><p>applied. Because of the multiple usage areas, there are many research efforts in face processing. Face detection is a challenging computer vision problem because of lighting conditions, a high degree of variability in size, shape, background, color, etc. To build fully </p><p>automated systems, robust and efficient face detection algorithms are required. </p><p> </p><p>Numerous techniques have been developed to detect faces in a single image; in this project we have used a classification-based face detection method using Gabor filter features. We have designed five frequencies corresponding to eight orientations channels for extracting facial features from local images. The feature vector based on Gabor filter is used as the input of the face/non-face classifier, which is a Support Vector Machine (SVM) on a reduced feature </p><p>subspace extracted by using principal component analysis (PCA). </p><p> </p><p>Experimental results show promising performance especially on single face images where 78% accuracy is achieved with 0 false acceptances.</p>

In this dissertation I explore the application of two novel modeling techniques for improving risk analysis of vector-borne disease and discuss their potential use in integrating environmental risk assessment that guides environmental and public health decisions. Techniques for analyzing risk have been considered inadequate due to a lack of understanding of the problem and an appropriate analytic-deliberative process clarifying the meaning of analytic findings and uncertainty (National Research Council (NRC), 1996; Peterman and Anderson, 1999). Thus, new integrative risk analysis tools are needed that are responsive to more complex environmental problems. In this work, I develop a qualitative community model that combines a conventional biomathematical model of vector-borne disease transmission with recent developments in community modeling. My procedure predicts the change in risk of vector-borne disease from press perturbations, a disturbance that results in a permanent change in a growth parameter. I also use a Relational Bayesian Modeling technique to exploit existing data to determine plausible mechanisms and geospatial and temporal patterns of disease spread. I apply these tools to Lyme disease and West Nile Encephalitis as examples of two different vector-borne diseases associated with complex ecological communities. Both the qualitative modeling and Bayesian methods provide an integrated risk analysis framework that identifies relationships important in the system and thus, guide the application of quantitative models or provide sufficient information for management decisions. / Graduation date: 2003

The image comparison operation ??sessing how well one image matches another ??rms a critical component of many image analysis systems and models of human visual processing. Two norms used commonly for this purpose are L1 and L2, which are specific instances of the Minkowski metric. However, there is often not a principled reason for selecting one norm over the other. One way to address this problem is by examining whether one metric better captures the perceptual notion of image similarity than the other. With this goal, we examined perceptual preferences for images retrieved on the basis of the L1 versus the L2 norm. These images were either small fragments without recognizable content, or larger patterns with recognizable content created via vector quantization. In both conditions the subjects showed a consistent preference for images matched using the L1 metric. These results suggest that, in the domain of natural images of the kind we have used, the L1 metric may better capture human notions of image similarity.

The importance of attribute vector ambiguity has been largely overlooked by the machine learning community. A pattern recognition problem can be solved in many ways within the scope of machine learning. Neural Networks, Decision Tree Algorithms such as C4.5, Bayesian Classifiers, and Instance Based Learning are the main algorithms. All listed solutions fail to address ambiguity in the attribute vector. The research reported shows, ignoring this ambiguity leads to problems of classifier scalability and issues with instance collection and aggregation. The Algorithm presented accounts for both ambiguity of the attribute vector and class label thus solving both issues of scalability and instance collection. The research also shows that when applied to sanitized data sets, suitable for traditional instance based learning, the presented algorithm performs equally as well.